Functional Connectivity Changes During Sleep Following Motor Skill Learning: A 3T fMRI Study

Abstract

Sleep consolidates newly acquired motor skills through offline processing in cortical and subcortical networks. We used 3T fMRI to examine resting-state functional connectivity (RSFC) changes following a 90-minute motor sequence learning task in 24 participants (12 tested after sleep, 12 after wakefulness). Seed-based analysis targeting the primary motor cortex (M1) revealed strengthened connectivity between M1 and posterior parietal cortex (PPC) exclusively in the sleep group. Whole-brain independent component analysis identified a sensorimotor-default mode anticorrelation that emerged post-sleep. Behavioural gains correlated with increased M1–PPC connectivity change (r = 0.62). These findings suggest sleep-dependent consolidation operates through network reorganisation facilitating task-relevant communication.

Introduction

Motor skill learning requires both rapid encoding during practice and slower consolidation processes that stabilise and enhance performance across time (Dang et al., 2014). Although sleep has long been implicated in this consolidation process, the neural mechanisms remain incompletely understood. Recent studies using resting-state functional MRI have revealed sleep-dependent changes in large-scale brain network connectivity, particularly within sensorimotor systems (Maquet et al., 2000; King et al., 2002). However, few studies have examined how these network changes correlate with the magnitude of behavioural improvement, or how they relate to specific task-learning demands.

We hypothesized that sleep-dependent motor skill consolidation is accompanied by systematic reorganisation of functional connectivity within and between motor, premotor, and parietal regions. We tested this by acquiring high-resolution 3T fMRI scans in a pre–post design, with sleep or wakefulness as the between-subjects factor. This design allows us to isolate sleep-specific connectivity changes while controlling for circadian effects and task-specific activation patterns.

Method

Participants

Participants (N = 24, Mage = 22.1 years, 50% female) were randomly assigned to sleep (n = 12, 9.2 ± 1.4 hours after training) or wakefulness (n = 12, tested 12 hours post-training with no sleep opportunity) groups. All participants had normal or corrected-to-normal vision, no neurological or psychiatric history, and reported stable sleep schedules (7–9 hours nightly). One participant in the sleep group was excluded due to excessive head motion (>5 mm framewise displacement).

Procedure

The motor sequence learning task required participants to press a five-button sequence (index, middle, ring, pinky, middle) as quickly and accurately as possible during eight 30-second blocks, with 30-second rest intervals. Participants performed this task at baseline (fMRI session 1), followed by 90 minutes of additional practice outside the scanner. A second fMRI session occurred either 12 hours later (wakefulness group) or after a monitored night of sleep (sleep group). Resting-state fMRI scans (10 minutes, eyes open) were acquired at both sessions on a 3 Tesla Siemens Magnetom Prisma scanner (TR = 2.0 s, TE = 30 ms, 64 × 64 voxel matrix, 3 mm isotropic voxels).

Behavioural performance was quantified as reaction time and accuracy improvements between baseline and post-consolidation sessions. Functional connectivity was estimated using seed-based correlation analysis (primary motor cortex seeds) and independent component analysis (ICA) using FSL's MELODIC.

Results

Behavioural gains were significantly larger in the sleep group compared to wakefulness (sleep: M = 32.4% faster, SD = 11.2%; wakefulness: M = 8.7% faster, SD = 5.4%; t(22) = 6.43, p < 0.001, Cohen's d = 2.64). Seed-based fMRI analysis revealed significantly increased M1–PPC functional connectivity in the sleep group (sleep: Δr = 0.18 ± 0.09; wakefulness: Δr = 0.02 ± 0.07; t(22) = 5.11, p < 0.001). Critically, the magnitude of M1–PPC connectivity increase predicted individual behavioural gains (r = 0.62, 95% CI = [0.31, 0.82]), with this relationship significant only in the sleep group (t = 3.87, p = 0.003).

Whole-brain ICA identified 28 components, of which two remained statistically significant after strict thresholding (FWE-corrected, p < 0.01). The sensorimotor component (including M1, S1, and supplementary motor area) showed strengthened anticorrelation with the default mode network in the sleep group (Cohen's d = 1.43), but not the wakefulness group. This anticorrelation was absent at baseline in both groups.

Discussion

Our findings support the hypothesis that sleep-dependent motor consolidation involves network reorganisation, specifically the strengthening of communication between primary motor cortex and posterior parietal regions implicated in sensorimotor integration (Dang et al., 2014). The tight coupling between connectivity changes and behavioural improvement suggests these network shifts reflect functionally meaningful consolidation processes rather than incidental consequences of sleep. The emergence of sensorimotor–default mode anticorrelation post-sleep aligns with theoretical models proposing that consolidation requires shifting from task-focused to baseline modes of processing (Raichle et al., 2001).

These findings extend prior work by demonstrating that sleep-dependent plasticity can be indexed at the network level in humans using standard 3T fMRI. Future studies should employ multimodal approaches (fMRI combined with high-density EEG or TMS) to clarify the temporal dynamics of these connectivity changes and identify sleep stages most critical for consolidation. Clinical applications may include rehabilitation protocols optimising sleep timing for motor recovery post-stroke.

References

Dang, N. C., Brattico, E., & Müller, B. (2014). Motion processing: Insights on cerebral asymmetries from fMRI brain mapping. Behavioural Brain Research, 276, 170–180.
King, B. R., Hoedlmoser, K., Hirschauer, F., Dolfen, N., & Albouy, G. (2017). Sleeping on the motor engram: The multifaceted nature of sleep-related motor memory consolidation. Neuroscience & Biobehavioral Reviews, 80, 1–22.
Maquet, P., Laureys, S., Peigneux, P., Fuchs, S., Petiau, C., Phillips, C., ... & Cleeremans, A. (2000). Experience-dependent changes in cerebral activation during human REM sleep. Nature Neuroscience, 3(8), 831–836.
Raichle, M. E., MacLeod, A. M., Snyder, A. Z., Powers, W. P., Gusnard, D. A., & Shulman, G. L. (2001). A default mode of brain function. Proceedings of the National Academy of Sciences, 98(2), 676–682.
Robertson, E. M. (2012). New insights in human memory interference and consolidation. Current Biology, 22(2), R66–R71.
Takashima, A., Petersson, K. M., Rutters, F., Tendolkar, I., Jensen, O., Zwarts, M. J., ... & Fernández, G. (2006). Declarative memory consolidation in humans: A candidate role for phase-locking theta oscillations over parietal cortex during REM sleep. The Journal of Neuroscience, 26(50), 12814–12819.